Film Dosimetry

CONTENTS

I. Introduction........................................................................................................................................................301
II. Examples of Photographic Film Dosimetry........................................................................................................301
III. Radiochromic Film Dosimetry............................................................................................................................310
References...................................................................................................................................................................329

1. INTRODUCTION
Most radiochromic systems are chemical radiation
Film dosimetry is attractive due to its high spatial sen- sors consisting of solid or liquid solutions of
resolu- tion, wide accessibility, and the flexibility to place colorless leuco dyes; these become colored without the
the film in humanoid phantoms. Also, the short need for deve- lopment when exposed to ionizing
measuring time and the fact that the film dosimetiy is radiation. Various radi- ochromic forms, such as thin
intrinsically two-dimen- sional and integrating in time films, thick films and gels, liquid .solutions, and liquid-
are appreciated. Film is potentially the ideal detector for core waveguides, have been in routine use for dosimetry
determining dose distribu- tion for dynamic beams and of ionizing radiation over a wide range of absorbed
for smdying the combination of stationary beams treated doses (10*2 to 10a Gy). Radiochromic film is used for
sequentially. Film dosimetry is widely used to obtain the general dosimetry of ionizing radiation in high-gradient
relative do,se distribution of elec- tron and photon beams areas ot electron and photon beams in a wide energy
in water, in plastics, and in inho- mogeneous phantom. range. The film allowing approximately tissue-
Film dosimetry in phantoms is advantageous because of equivalent dosimetry has been applied to mapping of
high spatial resolution, short treat- ment unit dose distribution in brachytherapy.
immobilization time, and 2D information. Radiochromic film(RCF) is of great interest as a
Modern film processing units improve the planar dosimeter for radiation oncology applications. It
reproduc- ibility and reliability of film dosimetiy and consists of a thin, radiosensitive, 7—23-pm thick,
make it an attractive method for many applications. Fa.st colorless leuco dye bonded to a 100-pm-thick mylar
film digitiz- ers connected to a computer, equipped with base. RCF turns deep blue in color upon irradiation.
proper eval- uation software, allow rapid and accurate RCF is approximately tis- sue-equivalent. GafChromic
analysis of large films in a short time. MD-55 is usable at doses from less than 1 Gy up to 12
The most common ,setup in relative dose Gy when measured at the wavelength of maximum
measurements with films is to sandwich a film within a sensitivity (676 nm) and up to about 500 Gy when
phantom of water- equivalent material with the film measured at a wavelength of low sensitivity. The
plane-parallel to the cen- tral axis of the radiation field. absorption spectra of GafChromic (GC) film contain
With the parallel geometry, two particular precautions two peaks with wavelengths in the range 610—680 rim
must be taken: there must be heavy pressure on the and, thus, the dose-response curve as mea- sured by an
phantom to avoid any air gap on either side of the film, optical densitometer or spectrophotometer will be highly
and there must be perfect alignment of the film edge dependent on the light source spectrum and sen- sor
with the surface of the phantom. In addi- tion, the material.
artifacts that result from a thin air layer trapped between
the packaging material, the paper spacer, and the fiIm
are responsible for the inaccurate dose measurement in II. EXAMPLES OF PHOTOGRAPHIC
the build-up region of the electron depth do.se. FILM DOSIMETRY
Conventional silver halide film has a highly nonlinear
photon energy response, especially at low energies. It The use of film as a dosimeter is still limited, due to the
has radiation interaction properties markedly different various difficulties associated with films such as energy
from those of tissue. Along with variations introduced dependence, film orientation, and sensitometric nonlinearity.
by the necessary post irradiation processing, thi,s type On the other hand, film is probably one of the best
of film is extremely difficult to use 1or accurate analytical detectors for studying spatial distribution of dose or
dosimetry. energy imparted. The dosimetric resolution is limited
only by the grain size
302 Radiation Dosimetry: instrumentation and Methods

FIG URE 5.1 Characteristic curves of the TVS films for photon beams. (From Reference [1]. With permission.)

and the size of the aperture of an optical densitometer. Figure 5.1 shows the variation of net optical density
Commercially available laser-scanning densitometers with dose for the TVS film for a range of photon energies
aiford film dosimetry with a resolution of a few microns.
from '*7Cs to 18 MV. The base optical density of the CEA
Unlike other detectors such as ionization chambers, diode
films i,s on the order of 0.06, compared to about 0.2 for
detectors, thermoluminescent detector,s (TLD,s), scintilla-
the Kodak films. As shown in Figure 5.1, the linear portion
tion detectors, and diamond detectors where dose informa-
of the char- acteristic curve covers a range of optical
tion mu.st be recorded before the readout is cleared for
density up to 4.3. Tlâs wide range of linearity oifers a
another irradiation, radiographic films allow repetitive
convenient means in percent depth dose and isodose
readouts and provide a permanent record of the dosimet-
measurements. It is inter- esting to note that the film is
ric measurements. Films may be customized in various
faster to rays than to bremsstrahlung x-rays from a linear
sizes and shapes to fit any dosimetric application. Due to
accelerator but is inde- pendent of energy for each type of
the relatively small thickness, a film can be treated very
radiation. This feature is particularly attractive for high-
close to a Bragg-Gray cavity. The physical flexibility of
energy x-rays, as one single sensitometric curve can be
a film is also suitable for the curved and cylindrical sur-
used for a wide range of x-ray energies. A straight line is
face do,se mapping when other detectors are impractical.
fitted to each of the characteristic curves for rays and x-
In relative dose measurements, the optical density may
rays with a regression coefficient of 0.999. For gamma
be taken as proportional to the dose without any correc-
rays, the line of regres- sion is [1]
tion, ,since the collisional stopping power ratio of emul-
sion to water varies slowly with electron energy. (OD)„,a„ = 0.054 07 X dose, (5.1)
High-energy photon and electron dosimetry was car-
ried out by Cheng and Das using CEA film. [1] The
while for x-rays, the line of regression is
packaging of the CEA films is distinctly dillerent from
that of the other films, in that each film is vacuum- sealed
in a shiny polyester-made, waterproof packet about 130 p. (OD)X ,a„ = 0.047 65 X dose (5.2)
thick, permitting film dosimetiy to be carried out even in
water phantom. Aside from the difference in packaging, Figure 5.2 compares the characteristic curves for the
the CEA films have a clear polyester film base as opposed CEA TVS film and the Kodak XV film for two x-ray
to a bluish-dye film base found in the Kodak Readypack energies, 4 and 10 MV. For the CEA TVS film, the linear
XV and XTL films. It i.s not clear if the dilference in film portion extends over the dose range up to 90 cGy,
base may have any effect on image quality. The CEA followed by an abrupt saturation above 90 cGy. The CEA
films come in two types: TLF (localization) and TVS TVS curves have a linear relationship, with a coefficient of
(verification), similar to the XTL and XV films of the regression very close to 1. For the Kodak XV film, on the
Kodak Readypack. other hand, the optical density varies curvilinearly with
Film Dosimetry 303

FIGURE 5.2 Comparison of the characteristic curves of CEA TVS film and Kodak XV film for different photon energies. (From
Reference [1]. With permission.)

FIGURE 5.3 Characteristic- curves of the CEA TLF film for photon energies 4—18 MV. For comparison, c-haracteristie curve of
Kodak XTL film is also shown for a 6-MV beam. (From Relerence [1]. With permission.)

dose up to 200 cGy. The characteristic curve of the Kodak CEA TVS film has a higher silver concentration (about 42
XV films exhibits a longer tail and a smaller gamma
g/m2) compared to regular x-ray films (about 7 g/m2).
compared to that of the CEA TVS film. The characteristic curves for the TLF film over the
The size of the silver halide crystal.s is generally photon energy range 4-18 MV are shown in Figure 5.3.
.smaller in the CEA TVS film than in the Kodak XV film. The film saturates at about 30 cGy, which is
Indeed, the silver halide crystals in the CEA TVS film are considerably higher than the Kodak Readypack XTL
of fairly uniform size and shape, re,sulting in a linear film. Unlike the CEA TVS film, the characteristic curves
characteri,stic curve. On the other hand, the ,silver halide for the CEA TLF films are slightly curvilinear over the
crystals of the Kodak XV film are of different sizes and range of dose up to 30 cGy. However, if the data is
shapes, with the largest more than 10 times bigger than the considered only in the range of 0—15 cGy, the curves
smallest, result- ing in a nonlinear characteristic curve. are all straight lines with a coefiicient of regression near
Furthermore, the unity. The CEA TLF film
304 Radiation Dosimetry: instrumentation and Methods

FIGURE 5.4 Film sensitivity dependence upon field size and upon depth. (From Reference [2]. With permission.)

is also energy-independent within the experimental The effects of foil-to-film separation distance and
uncertainty. foil thickness were investigated in order to obtain a
A method of film dosimetry for high-energy photon single optimum distance-thickness combination, and the
beams was proposed by Burch et al. [2] which reduced results are presented in Figures 5.5 and 5.6. In Figures
the required film calibration exposures to a set o1 films 5.5 and 5.6, the dose was calculated using data from the
obtained for a small radiation field size and shallow 6-cm X 6-cm calibration films. Film/foil separation
depth (6 cm X 6 cm at 5-cm depth). It involves distances of 0, 0.6, 1.2, and 1.9 cm and lead foil
modification of a compres- sion-type polystyrene film thicknesses of 0, 0.15, 0.30, 0.46, and 0.76 mm were
phantom to include thin lead foil,s parallel to the vertical included in the investiga- tion. At 0-cm film/foil
film plane at approximately 1 cm thorn both side,s of the separation distance, the curve shows the effect of
film emulsion. The foil,s act as high atomic number filters electrons coming from the lead due to interactions
which remove low-energy Compton scatter photons that within the foil. This intensification effect exaggerates
otherwise would cause the film sensi- tivity to change the shape of the depth-dose curve and the non- uniform
with field size and depth. film/foil contact is apparent in the data (Figure 5.5). At 0.6
High-energy photon beams used in radiation oncology cm the scattered electrons are absorbed in the inter-
are considered to interact primarily by the Compton vening polystyrene. Further increase in film/foil
scat- tering process with tissue. However, when film is separation distance produced only minor changes in the
placed in a tissue-equivalent material, photoelectric calculated dose curve. Figure 5.6 shows the effect of
interactions associated with the silver atoms in the changing foil thickness. A single thickness of 0.15 mm
emul,sion cause the film to over-respond relative to the
cau,ses a signif- icant decrease in the calculated dose,
tissue-equivalent mate- rial. For the low-energy
with only subtle changers as additional layers are added.
photon,s, film do.se may be a,s much as 25 times the
The best match to ion chamber percentage depth-dose
ti,ssue dose at the same physical location.
data was observed for lead foil thickness of 0.46 mm
The results for the 6-cm X 6-cm and 25-cm X 25-
with a 1.2 cm film/foil separation distance.
cm field,s shown in Figure 5.4 empha,size the The use of radiographic film for the dosimetry of large
importance of the film sensitivity as a function of field (i.e., >10 cm X 10 cm) photon radiotherapy beams is
size and depth. As the field size was increased, the dose complicated, due to a dependence of film emulsion sen-
required to produce a given density on the film was sitivity on depth within the phantom. This dependence
reduced. For sizes smaller than 10 cm X J0 cm, the is caused by a relative increase in depth in the
sensitivity change i,s small (not shown); and 1or size,s population of lower-energy, scattered photons in the
le,ss than 6 cm X 6 cm, no change in sensitivity occurs. spectrum and the subsequent photoelectric absorption
For small fields with les,s scatter, the change in of these photons by the film emulsion. This effect
.sen,sitivity with depth is not apparent. becomes most pro- nounced with increases in the
photon population in the
Film Dosimetry 305

FIG URE 5.5 Lead foils placed adjacent to the film show an exaggerated response at shallow depth and a wavy appearance due to
undulation in their surface. Each c-urve is normalized to fi-em depth. (From Reference [2]. With permission.)

FIGURE- 5.6 The ettect ‹ai foil thickness on the shape of the depth-dose curve. (From Rcierenc-e [2]. With pcrmissit›n.)

energy region below approximately 400 keV, where the

ma.ss attenuation coefficient of emulsion increases Kodak X-Omat V film from a single batch (#194 05 2)
rapidly, approximately as I/(/iv)". As the low-energy was sealed within a light-tight Solid Water (Gammex
spectral component increases, the tissue equivalence of RMI, Inc.) cassette. The cassette consists of two 2-cm-
film emul- sion diverges from that of tissue-equivalent thick slabs of solid water sealed around three edges by
materials, due to an approximate Z dependence of the nylon screws and a rubber O-ring. [3]
photoelectric interaction cross section. [3] To minimize possible variations due to film-
In order to examine possible dependencies of emul- processing conditions, a Kodak X-Omat RP processor
sion exposed by radiosurgical beams, a series of was used by Robar and Clark for which the throughput
sensito- metric curves was established for a range of is very high and quality assurance is performed daily.
depths and field sizes. For each exposure a 10-in X Developer tem- perature fluctuated by less than *0.5°F
12-in sheet of between pro- cessing sessions. Optical density was
measured using a
306 Radiation Dosimetry: instrumentation and Methods

scanning densitometer. Base-plus-log optical density was

subtracted from scanned optical densities for each film.
In order to relate optical densities to absolute doses,
the dose to water corresponding to each of these depths
and for each MU .setting was calculated lrom

MUPDD(d, A)S,(A)
D(d, A) 100 (5.3)

where D{d, A) is the dose at depth d for field .size A at the

phantom surface, MU is the number of monitor units
given, PDD(d, A) is the percent depth dose and S,(A) is
the total scatter factor. Both S,(A) and PDD(d, A) were
measured in a water phantom using a p-type silicon elec-
tron diode .
Figure 5.7a shows the sensitometric curves obtained
for the large (20-cm X 20-cm) photon beam at depths of
1.0 cm,10.0 cm, and 20.0 cm. For each depth a curve was
fitted to the measured sensitometric data using the single-
target/single hit equation

D
OD OD.saw(1 — 10 ° ) (5.4)

where OD and D are the measured optical density and

given dose, respectively. The ,saturation density of the
film, OD a„ was estimated by delivering a large dose (500
cGy) to a film in phantom, and it was held constant in the
fitting
algorithm. The fitting parameter a represents emulsion
sensitivity and was allowed to vary. For this large field,
the curves diverge markedly, indicating that using a sen-
sitometric curve that is not depth-specific would introduce
significant error in converting optical density to dose. In
order to minimize this error to below approximately 109c,
for example, it is necessary to confine the dose range to
less than 50 cGy. In contrast, Figure 5.7b shows sensito-
metric curves corresponding to the same depths for the
radiosurgical (2.5-cm-diameter) field. For this small field,
the curves agree to within the reproducibility of film FIGURE 5.7 The sensitometrie curves established experimen-
devel- opment and scanning. By determining the value of tally for (a) the 20-cm X 20-cm radiotherapy field and (b)
a for each curve, film sensitivity was obtained as a the 2.5-cm-diameter radiosurgical field, shown for depth in
function of depth, as illustrated for the 20-cm X 20-cm solid water phantom of 1.0 cm, 10.0 cm, and 20.0 cm. (From
and 2.5-cm- diameter fields (Figure 5.8). A,s expected Refer- ence [3]. With permission.)
from the dispar- ity in the sensitometric curves in
Figure 5.7a, the film sensitivity for the large (20-cm X number of the phantom material and the physical size and
20-cm) field increases systematically with depth. While orientation of the dosimeter. Significant differences were
small fluctuation of the values of n is apparent for the observed by El-Khatib et al. between the dose measure-
radiosurgical field, no systematic variation of sensitivity ments within the inhomogeneity. [4] These differences
with depth is apparent. Film dosimetry is most problematic were influenced by the type and orientation of the dosim-
for lower-energy photon beams (e.g., ^0Co and 4-MV) eter in addition to the properties of the heterogeneity.
due to their lower primary-to-,scatter ratio. Figure To investigate how the differences in dose measure-
5.7 ,shows the importance of corrections in converting ments are related to the dosimeter configuration, the EGS4
scanned optical density to dose Monte Carlo system, together with the PRESTA
even for the 6-MV beam. algorithm, was used to model the experimental condition,s.
Film is often used for dose measurement,s in hetero- The per- centage depth doses measured in the phantom
geneous phantoms. In those situations perturbations are containing 1-cm bone are shown in Figure 5.9 and are
produced which are related to the density and atomic compared to
Film Dosimetry 307

FIGURE- 5.8 The curve-fitting parameter n, representing emulsion sensitivity, as a function of depth for the 20-cm X 20-
cm radiotherapy field and the 2.ñ-cm-diameter radiosurgical field. (From Reference [3]. With permission.)

FIGURE- 5.9 The percentage depth doses measured in a polystyrene phantom containing 1-cm hard bone at 1-cm depth using
(A) film in parallel orientation to the beam, (B) film in perpendicular orientation to the beam, and (C) TLD are Shown and compared
to the PDD measured in a homogeneous water phantom for electron beams of nominal energy (a) 12 MeV and (b) 20 MeV.
(From Reference [4]. With permission.)

the depth doses measured in the homogeneous water phan- for the two film orientations is observed for the phantom
tom. All film measurements were done with the film in the containing the cork (Figure 5.10). The PDD is unaffected
ready pack with the inner paper removed. The doses were in the initial 4 cm of polystyrene and there is greater
normalized to the maximum dose in the homogeneous penetration of the beam in cork. Within the cork the dif-
polystyrene phantom, measured with whatever dosimeter ference measured with the film in different orientations is
was used to measure the doses at depth. The doses within as much as 69a, and the TLD measurements are lower than
the bone represent dose-to-unit density material within the the measurements with film. These greater discrepancies
bone rather than dose to the bone itsell, since a correction are attributed to the fact that the inhomogeneity i.s much
by ratio of collisional stopping power,s was not done. At larger and lies in a region of high do,se gradient.
several cm beyond the bone, the PDD measured with all A method for the creation of a complete 3D dose
dosimeter configurations are identical. The opposite effect distribution from measured data was discussed by
308 Radiation Dosimetry: instrumentation and Methods

FIGURE 5.10 The perc-entage depth doses measured in a polystyrene phantom containing 6-cm cork at 4-cm depth using (A) fi1m in
parallel orientation to the beam, (B) film in perpendicular orientation to the beam, and (C) TLD are shown and compared to the PDD
measured in a homogeneous water phantom for electron beams of nominal energy (a) 12 MeV and (b) 20 MeV. (From Reference [4].
With permission.)

FIGURE- 5.1 4 Film sensitometric curves derived from four sets of BEV films placed at depths of 3, 5, 10, arid 20 cm, respectively.
(From Reference [ñ]. With permission.)

Stern et a1. [5] It involved the measurement with film of

the dose distribution in a series of “beam’s-eye-view” When necessary, the number of pixels was reduced using
(BEV) planes (planes perpendicular to the beam central nearest-neighbor sampling. The images were entered into
axis at different depths). Film measurements were made the planning system and aligned with the planning system
using Kodak Readypack XV-2 film (Eastman Kodak Co., representation of the beam by translating and rotating the
Rochester, NY) sandwiched between sheet,s of water- coordinate system of the displayed film image. Optical
equivalent solid. Exposed films were digitized using a density values were converted to dose values using the
la.ser digitizer with spot .size .set to 0.42 mm and pixel appropriate measured film sensitometric curve.
size 0.45 mm. The sensitometric curves obtained at Your different
Film image data files were next converted into the depths ranging from d to 20 cm are plotted vs. absolute
planning system's standard grayscale image file format. dose in Figure 5.11. The largest variation in dose for a
given optical density determined from this set of curves is
Film Dosimetry 309

FIGURE 5.12 Relative sensitometric curve for three photon beam qualities (°’Co, 6-MV x-rays, 18-MV x-rays). (From Reference [6].
With permission.)

2.39c, at an optical density of approximately 2.3. Variation No increase is observed in the fog value for un-
is less than ldc over most of the measured dose irradiated films after mailing or storing for more than
range. This demonstrates that a single sensitometric 75 days; the average reading is 0.J6 * 0.0J for the
curve mea- sured at one depth can be accurately used two series of films, either processed immediately after
for optical density-to-dose conversion of BEV films at irradi- ation or processed after mailing.
all depths for the experimental ,situation used. Fading is expressed as the ratio of the optical
A feasibility study for mailed film dosimetry has been density of a faded film and the optical density of the
performed by Novomy et al. [6] The fading of the reference film, as a function of the time t between film
latent image before film processing is only 39a per irradiation and film processing. The interval t is equal
month and the normalized sensitometric curve is not to 0 for the reference film. The dependence of fading
modified after a period of 51 days between irradiation on time t, for films stored in the laboratoiy, is
and processing. All film experiments have been presented in Figure 5.13. A straight line is fitted
performed with Kodak Readypack paper envelope through the measured data using the least- squares
films (25.4 cm X 30.5 cm). All films were processed method. A decrease of about 39a in optical density is
with an automatic processing unit, either an Agfa observed per month of delay between film irradiation
Gevamatic 1100 (processing time 90 s and temperature 35 and film processing (corresponding to a given dose of 0.45
* 1°C) or a modern AGFA Curix 260 (processing time Gy and optical density of about 1.3). A smaller
90 s and temperature 35 * 1°C). No dependence on decrease of about 1.89c per month in optical density is
the photon energy has been observed in the shape of observed for films irradiated with dose of 1.5 Gy
the normalized sensitometric curve for ^"Co, 6-MV, and (corresponding to optical density 2.7). These
18-MV beams (Figure 5.12). decreases correspond to esti- mated variations on dose of
3.89c and 3.49c, respectively.
34 0 Radiation Dosimetry: instrumentation and Methods

FIGURE 5.13 Decay with time of optic-a1 density (fading) of Kodak X-omat V films irradiated with 0.4fi Gy or 1.ñ Gy. The error
bars represent one Staiidard deviation of total uncertainty. (From Reference [6]. With permission).

III. RADIOCHROMIC FILM DOSIMETRY broadband light source of a conventional densitometer is

The introduction of radiochromic dye-cyanide films has attenuated only in the red part of the spectrum. This lead,s
resolved some of the problems experienced with conven- to an overall reduction in the light intensity which is quite
tional radiation dosimeters. The high spatial resolution small and, thus, the resulting measurements have a low
and low spectral sensitivity of radiochromic films make signal-to-noise ratio. Densitometers employing a He-Ne
the dosimeter ideal for the measurement of dose distri- laser, which emits light at 633 rim, close to the peak of
butions in regions of high dose gradient in radiation fields. about 660 nm in the absorption spectrum of the film, have
The three radiochromic films mo,st commonly used in been used. This improved the sensitivity and the linearity
medical applications have Been GafChromic MD-55 of the dose response. These densitometers can make mea-
(Nuclear Associates Model No. 37-041), GafChromic surements with resolutions up to 0.3 pm. [7]
HD-810 (Nuclear Associates Model 37-040), and The images produced by the scanner were loaded into
GafChromic DM-1260. The MD-55 films are available the software package MATLAB (The Mathworks, Inc.,
in 5“ X 5” clear sheets and are suitable for dose mea- version 4.2c). This allowed for the representation of the
surements in the range of 10—100 Gy. Model HD-810 image as a matrix which may be manipulated for the
films were available until recently in 8“ X 10” sheets and purpo,ses of image enhancement, enlargement, and map-
were used for measurements of absorbed dose in the ping of isodoses or depth doses.
range of 50—2500 Gy. Model DM-1260 films were avail- A calibration to convert from raw scanner signal to
able in 5“ X 5” rolls and also were used for dose mapping dose was achieved by placing 12-mm X 30-mm pieces of
in the range of 50—2500 Gy. film on the surface of a 10-cm-thick Perspex phantom of
Radiochromic film ha,s ,significant advantages over dimensions 20 cm x 20 cm and irradiating them to doses
silver halide film: it has a relatively flat energy response; in the range 0—95 Gy.
it is self-developing so it eliminates variations introduced The net optical density, OD, of a point on the film is
by the processing step; it is insensitive to visible light, given by the equation
allowing for ease of handling; and the film is fabricated
from low-atomic-number materials, so it does not perturb OD log ( S (5.5)
the radiation beam to the same degree as silver halide film.
The response of the film has been shown to be
independent
of dose rate. The uniformity of the film means that no where S0 is the “dark current,” i.e., the scanner signal for
background measurements need to be taken prior to irra- an unexposed film, and S is the scanner signal for the fi1m
diation. The film is .sen.sitive to ultraviolet light with at the point of interest. The optical density depends on the
photon energies greater than 4 eV. [7] dose D in the following way:
Unexposed radiochromic film is transparent with a
light blue hue, while film which has been exposed to S)
log,
0 aD‘ (5.6)
ionizing radiation is dark blue. The blue coloration means
that the
Film Dosimetry 311

FIGURE- 5.14 Curve showing the dependence of optical density on dose. The data pt›ints correspond to the t›ptio-al density t›f films
uniformly irradiated to doses in the range 0—95 Gy, with 55-kV x-rays at the surface of a Perspex phantom. The error bars demonstrate
the level of uncertainty in the optical density, based on the uncertainty in measurements of scanner signal. The solid line shows
how the model represented by Equation (5.6) fits the experimental data. (From Reference [7]. With permission.)

where o and n are constants. These constant,s are deter-

mined by taking the log o1 Equation (5.6), resulting in

0
1n[lo
S In(o) + «In(D) (5.7)

A plot of ln [log
)] against ln(D) produces a
straight line with slope n, intercept In(a), and linear
regression coefficient r — 0.999. A least-squares linear fit
to the calibration data yields the constant,s a = 0.719 and
a = 0.01 50. Equation (5.6) is solved for dose and
becomes, after substitution of the constants,

D 5.56log„($) ”’ (5.8)

The dependence of optical density on dose is shown in

Figure 5.14, with Equation (5.6) represented by the solid
line. It was shown that the optical density can change by
169c in the 24 h immediately after irradiation, but in the
following 24 h the color change is less than 1'fa. [7]
GafChromic*M Dosimetry Media offers advances in
high-dose radiation dosimetry and high-resolution radio-
graphy for gamma radiation and electrons. [S] The film
consists of either small 1.2-cm X 6-cm strips as individual
dosimeters or, for the purpose of large-scale imaging, FIGURE 5.15 Absorption spectra of GafChromie*^ Dosimetry
0.13-m X 15-m roll.s of coated po1ye.ster film, or other Media, un-irradiated and irradiated to different doses.
sizes up to a meter wide and 50 m long. “The sensitive (From Reference [8]. With permissiran.)
coating is a proprietary sensor layer 6 pm thick, and
ation-induced absorption spectra are shown in Figure 5.15.
the base is a 100-pm-thick clear polyester film.” The film
Doses down to less than 1 Gy can readily be measured by
is colorles,s and “grainles,s” in that it,s radiographic image
using a special form of the GafChromic T* Dosimetry
gives a spatial resolution of >1200 lines/mm. Typical radi-
Media.
312 Radiation Dosimetry: instrumentation and Methods

The increases in absorbance (AA) at four wavelengths Figure 5.18 shows hypsochromic shift with storage
are presented in Figure 5.16 as a function of the absorbed
time for large doses, where measurements have to be made
dose in water. The four curves are approximately linear
in the near ultraviolet or near-infrared because of
functions within the dose range up to 0.5 kGy, but they
unmeasurably high absorbance values in the visible
become nonlinear over the range lrom 0.5 to 25 kGy
spectrum. The readings on the long wavelength side are
(Figure 5.17; McLaughlin ct a1. [8]).
more stable than those in the blue and UV part of the
spectrum. Long-term stability of the readings of the films
at all wavelengths after a one- day initial storage period
remains within 49c, when the stor- age is under controlled
laboratoiy conditions. This long-term stability is observed
at all doses from 10 to 1200 Gy.
Figure 5.19 reveals that there are slight shifts to
shorter wavelengths of the absorption band maxima as the
absorbed dose increases.
Vatnitsky et al. [9] have studied the feasibility of
radiochromic film for dosimetry verification of proton
Bragg peak stereotactic radiosurgery with multiple beams.
High-,sensitivity MD-55 radiochromic film was calibrated
for proton beam irradiation, and the RIT 113 system was
employed for film evaluation.
In order to take into account geometry-related uncer-
tainties, the aperture of each proton beam should include
a margin around the target, and the bolus should be
“expanded” or “smeared.” The choice of margins and
bolus smear depends on the treatment technique and varies
FIGURE- 5.16 Curves representing the response of from center to center.
GafChro- mic'^ Dosimetry Media to gamma radiation from High-sensitivity MD-55 film consists of a 30-pm
a ^"Co souro-e, when the absorbance values are measured polymer sensor layer coated on a 100 pm polyester base
at the indicated wavelengths with a spectrophotometer, as and sandwiched between two plastic sheets. 5” X 5“ film
a func- tion of the absorbed dose in water. The vertical sheets were used and stored in black light-tight plastic
bars on each data point indicate the estimated limits of envelopes at controlled room temperature (22°C) and
random uncertainty at a 959c confidencc level (2s). (From humidity (50' a). Unpacking, cutting the film to the desired
Reference [fi]. with permission.) size, irradiation, and evaluation of the films were per-
formed under normal room light. Un-irradiated films were

FIGURE- 5.17 Response curve for gamma radiation doses up to 25 kGy, when measured with a spectrophotometer at 400-
rim wavelength. The vertical bars indicate the 2o- (959a confidence level) uncertainty to the random values of net absorbance at each
dose (in water), and the horizontal bars give the corresponding 2s uncertainty limits of the absorbed dose readings. (From
Reference [8]. With permision.)
Film Dosimetry 313

FIGURE- 5.18 Changes in the absorption spectra of GafChromic** Dosimetry Media exposed to high doses of gamma radiation 10,
20, and 30 kGy, and measured at different times after irradiation. (From Reference [8]. With permission.)

handled under the same conditions to control the back-

ground optical density. Evaluation of all films was per-
formed 24 h after irradiation.
Irradiated film,s were process.sed with the RIT 113
film dosimetry system (RIT Inc., Denver, Colorado). The
sys- tem configuration includes a digitizer (Argus II) and
soft- ware for film analysis, which consists of an image-
pro- cessing platform and various application modules.
The Argus II digitizer uses a daylight fluorescent light
lamp and CCD (charge coupled device). The CCD's linear
sen- sor has a maximum resolution o1 600 horizontal X
1200 vertical dpi (dots per inch). The RIT 113 system is
typically used in the following manner. First, calibration
films exposed to various doses are scanned in the film
digitizer to convert the film to a sequence of pixel,s who,se
value.s repre.sent the optical density at each point on the
film. A calibration file is generated to provide the system
with the information nec- essary to convert optical
densities on the film to radiation dose levels. Second,
subsequent films are exposed to the radiation beam, to
qualify the radiation beam characteris- tics. These films
are then digitized and analyzed by the software. [9] Figure
5.20 show,s the calibration dose response of the MD-55 FIGURE 5.19 Absorption spectra of Gafchromie*^ Dosimetry
film to the irradiation at the center of the spread-out Bragg Media, unirradiated and irradiated to the two indicated doses, at
peak (SOBP) for two beam ener- gies as a function of the indicated temperatures. (From Reference [8]. With permission.)
absorbed dose to water.
plastic spacer and aligning the detector block according
For the five-beam irradiation ,study, the detector block
to labeled marLs on the water phantom. The positioning
with radiochromic films was installed in the water phan-
uncertainty of the isocenter projection on each film was
tom so that the intersections between the film planes and
estimated to be less than 1 mm.
CT planes were parallel to the vertical CT axis. Potential
Figure 5.21 shows calculated and measured dose pro-
errors associated with the angulation of the film were
files for the five-beam composite plan in the isocenter
minimized by cutting each film exactly to the shape of the
344 Radiation Dosimetry: instrumentation and Methods

FIGURE- 5.20 Calibration response of MD-55 film to proton irradiation at the center of the SOBP. (From Reference [9]. With
permission.)
plane (A) and in the plane 3 cm superior to the isocenter
(B). The multiple-beam profiles were normalized to a fluctuation. RCF response was found to deviate signifi-
dose of 70 Gy. The compari,son presented in Figure 5.21 cantly from linearity at low doses (409a change in net
dem- onstrates that the treatment plan provided adequate OD/Gy from 1 to 30 Gy).
target coverage and that the dose delivered to the The film consisted of a thin, radiosensitive (7—23
reference point (center of the target at isocenter plane) pm thick) colorless leuco dye bonded to a 100-pm-thick
was close to the prescribed dose within accuracy of film Mylar base. RCF films are colorless before irradiation
do.simetry (*59c). The absolute dose difference between and turn a deep blue color upon irradiation without
calculated and mea- sured dose values over the target physical, chemical, or thermal processing. RCF is
irradiated with multiple beams was also found to be approximately tissue-equivalent, integrates
within the uncertainty of the radiochromic film simultaneously at all mea- surement points, and has a
dosimetry (*59a, one standard devi- ation) for the high spatial resolution (> 1200 lines/mm). It shows a
stable, reproducible response if pro- tected from UV
isocenter plane and off-center planes. The observed
light and unstable temperatures and humidity. Since
difference between measured and calculated profile.s
radiochromic dye is an aromatic hydro- carbon, like
can be explained by the dependence o1 the radi-
plastic scintillator, its deviation from tissue- equivalent
ochromic film sensitivity at a very steep distal edge of
energy response is much smaller compared to silicon
the Bragg peak.
Zhu et al. [10] have studied the uniformity, linearity, diode detectors and has the same magnitude (but
and reproducibility of a commercially ,supplied radio- opposite direction) as that of TLD.
chromic film (RCF) system (Model MD-55). Optical To suppre.ss the effects of nonuniform RCF
den- sity (OD) distributions were measured by a response, a double-exposure densitometry technique
helium-neon scanning laser (633 rim) 2D densitometer was used. Before use, each film was first given a
and also with a manual densitometer. All film strips uniform dose of 20 Gy or 40 Gy, and fiducial marks
showed S—159c vari- ations in OD values, independent were placed on two opposite corners of the film using a
of densitometry tech- nique, which are evidently due to punch with a diameter of 600 pm. After a time interval
nonuniform dispersal of the sensor medium. A double ranging from 48 to 96 h from irradiation, the fi1m was
expo,sure technique was used to solve this problem. The then scanned and the result- ant optical density
film is fir,st expo,sed to a uniform beam, which defines a distribution, OD (i,j), was used to form a pixel-by-pixel
pixel-by-pixel nonuniformity cor- rection matrix. The sensitivity correction matrix, f(i,y).
film is then exposed to the unknown dose distribution, OD (i, j)
re,scanned, and the net OD at each pixel corrected for f '‘• 1) ( OD [i, j)) (5.9)
nonuniformity. The double exposure tech- nique reduces
OD/unit dose variation to a 2—59a random where i,y are the X and Y position indices and (OD (i,
j)) is the mean OD over the entire film. This pre-exposed
film
Film Dosimetry

aligning the fiducial marks, and then calculates and dis-

plays the net optical density distribution, kOD(i, j) cor-
rected for pixel-to-pixel sensitivity variations:

ODz(!‘›
b OD(i, ;) '(' ) 5.10)

The corrected net optical density, bOD(i, j), is taken

to be the response of the detector to the unknown radiation
field. Pixel-by-pixel correction for nonuniformity of film
response was performed using the original OD matrix
consisting of 50-pm pixels. Films were always scanned
with the sensitive side of the RCF emulsion facing the
ceiling to prevent the two OD images from being flipped.
[10]
The response of RCF with respect to absorbed dose in
a 6-MV x-ray beam for both single- and double-exposure
techniques was studied by Zhu et a1. using the same films
and OD distributions as were used for interfilm reproduc-
ibility analysis. For each of the dose groups, the mean of
the average optical ID density of each film was taken,
yielding the final mean optical density (b OD) corre-
sponding to the net dose, ID. The deviation of RCF
response from linearity was quantitated by means of a
relative linearity correction factor F „, defined as

F„,(b OD) — [ (b OD)/bD] At each dose level

[ (b OD /bD] At a reference of 30 Gy
(5.11)

Zhu et a1. chose the 30-Gy dose level for normalizing

Jt;, to unity because it represents the center of the dose
range typically used in their RCF dose measurements;
however, its choice is completely arbitrary and affects
neither the accuracy nor precision of subsequent dose
FIGURE- 5.24 A comparison of measured and planned dose measurements. The linearity correction factor of RCF
profiles for five-beam irradiation. Vertical lines represent bound- response using the Macbeth manual densitometer was
aries of a sliced polystyrene cylinder with films in the defined as
detector block. (A), a c-omparison of measured and planned
dose profiles for five-beam irradiation, isocenter plane. (B), a
comparison of
measured and planned dose profiles for five-beam irradiation, [ ( OD)/D At each dose level
CT plane, 3 cm superior to the isocenter. (From Reference F (A OD)
[ ( OD)/D] At a reference of 30 Gy
[9]. With permission.)
(5.1 2)
is then exposed to the unknown dose distribution, held for
an interval of 48—108 h, and scanned, yielding the cumu- where (b OD) represents the single-exposure OD aver-
lative optical density distribution, OD2(i,y).The time inter- aged over manual readings measured over a 3 X 3-mm
val between the OD scan and exposure to the unknown grid from each film averaged over the four films in each
radiation field was highly variable, ranging from a few dose group with background density subtracted. For doses
hours to several weeks. Both OD and OD2 images are then over 20 Gy, the cumulative OD2 densities were measured
transferred to the graphics work.station. There, a .software and plotted as a function of Do .To measure the sensito-
package displays the two image,s, allows corresponding metric curve at lower doses, a series of unexposed films
fiducial marks on the two films to be identified, correlates were irradiated to the desired doses and readout as
the double- and single-exposed images pixel-by-pixel by described above.
346 Radiation Dosimetry: instrumentation and Methods

Figure 5.22 shows a typical OD profile through the

center of a 12-cm-long strip, illustrating the sinusoidal
behavior of its nonuniform response. The upper panel
shows the profile resulting from a uniform dose of 40 Gy,
while the center panel illustrate,s the elfect of an
additional 20-Gy exposure. This demon,strate,s that the
nonuniformity is a multiplicative phenomenon,
independent of the dose delivered. Finally, the lower panel
of Figure 5.22 illus- trates that nonuniform film response
is dramatically reduced by the double-exposure method.
An 119c system- atic variation of film response is reduced
to a random OD fluctuation, which is less than 5'fa.
Figure 5.23 shows the results of interfilm reproduc-
ibility for both the single and double-exposure techniques.
For both techniques, the average OD in the central 5 X 5
block of 50-pm pixels was extracted from each film and
then averaged over the four films in each do,se group for
doses ranging from 1 Gy 230 Gy. For the double-exposure
technique, an initial 20-Gy exposure was used to define
the relative pixel sensitivity correction. The graphs also
show fi1m response as a function of absorbed dose. For
the single-exposure technique, the relative standard devi-
ation is as large as 11'fa at low doses and stabilizes at a
value of 49c for doses in excess of 30 Gy. For the double-
exposure technique, the relative standard deviation is as
high as 99c at low doses and rapidly falls to a plateau of
29c at 30 Gy. This result indicates that for Your repeated
readings using separately exposed films, the double-
exposure technique should be able to characterize radiation
fields with an exper- imental precision of *29c (959c
confidence interval) when exposed to doses of at lea.st 20
Gy.
GafChromic MD-55 was irradiated with Co prays by
Klassen et a1. [11] A double irradiation method was used
in which a dosimeter is given an unknown dose followed
by a known, calibration dose. It was found that the
measured optical den,sity of GafChromic MD-55, as
currently fab- ricated, is affected by the polarization of the
analyzing light, an important consideration when using
Ga?Chromic MD-55 as a precision do,simeter.
The thin, radiosensitive layers in GafChromic MD-55
are made of a colorless gel containing polycrystalline
substituted—diacetylene monomer, which undergoe.s par-
tial polymerization when irradiated with ionizing radiation
such that the blue color of the polymer becomes progres-
sively darker as the dose increases. The DOD due to the FIGURE 5.22 Central line OD profiles from a single film.
polymer increases roughly linearly with absorbed dose. Graph (a) shows the results of a single 40-Gy exposure.
Figure 5.24a show,s the optical absorption ,spectrum of The middle figure, graph (b), shows the profile resulting
GafChromic MD-55 from 600 nm to 700 nm, before and from an additional dose of 20 Gy. The lower figure gives
after a dose of 6 Gy. Figure 5.24b, with it,s expanded scale the net OD profile [derived by pixel-by-pixel subtraction of
graph (a) from graph (h)], with (dashed line) and without
of OD and wavelength, ,shows the spectrum of an un-
(solid line) the pixel- by-pixel uniformity correction derived
irradiated film from 700 nm to 720 nm taken with a band-
from the initial 40-Gy exposure. (From Reference [10]. With
pass of 1.0 nm. The oscillations in the OD are interference permissican.)
fringes which constitute a channel .spectrum. [11]
Some of the apparent absorption is acmally a loss of for reflective losses,
light due to reflection, not absorption. Using the equation
Film Dosimetry 3J 7

FIGURE 5.23 Dose response c-urve of RCF to 6-MV x-rays at doses ranging from 1 Gy to 230 Gy for single- and double-
exposure techniques plotted on the left-hand scale. Interfilm reproducibility (relative standard deviation of the central effective
pixel readings from the four film pieces) is plotted on the right-hand axis of the graph. (From Reference [10]. With permission.)

FIGURE- 5.24 (a) The ‹ ptical absorption spectrum ot un-irradiated GatChromic MD-55 (lcawer spectrum) and GatChromic MD-.55
several weeks alter a dose of about 6 Gy (upper spectrum). The bandpass used was 3.5 nm. Some useful wavelengths are indicated.
(b) The spec-trum of un-irradiated GafChromie MD-55, taken using a bandpass of 1 nm and displayed at a higher sensitivity than in
(a) in order to demonstrate the periodic fluctuations due to interference fringes (channel spectrum). (From Reference [11]. With
permission.)

where fi is the reflectivity, 100a is the percent reflection, By returning to the initial temperature several hours later,
qp d„p is the refractive index of the medium (taken to be
it was found that the OD did not change irreversibly
I.5 for Mylar), and q„, is the refractive index of air (I .0).
during the measurements. The results, displayed for
Klassen et al. calculated that the reflection from each
673, 674, 675, 676, 677, and 678 nm in Figure
Mylar/air surface leads to a 4.09o light loss, or 8.09o from
5.25, ,show that at 22°C the temperature dependence is
the two ,such .surfaces, which corre,sponds to an OD o1
smallest at 675 nm, only 0.0002 OD units for a 0.1
0.036. There are seven 1ayer,s in GafChromic MD-55,
degree change in temperature between 20°C and 24°C.
meaning eight reflecting surfaces. At 24°C, a wave- length of 674 nm has the smallest
The temperature dependence o1 OD at 676 nm was dependence of OD on temperature. Figure 5.25 shows
measured in two studie,s u,sing six dosimeters that had that poor control of the spectrophotometer temperature
received 0, 1.0, 3.5, 6.9, and 14.5 Gy. Readings were can be countered by selecting the wavelength with the
taken at seven temperatures between 18.8 and 28.1°C. smallest temperature dependence.
8 Radiation Dosimetry: instrumentation and Methods

FIGURE 5.25 A plot of the OD of a single dosimeter, whic-h had received 6.9 Gy as read at a variety of temperatures and
wavelengths. The measurements were made several months after the irradiation. (From Reference [11]. With permisSiOIl.)

FIGURE 5.26 The rate of change of optional density of un-irradiated and irradiated GafChromic MD-55 dosimeters where DOD
is expressed in terms of dose (assuming that 1 Gy = 0.11 OD units). Time zero for the irradiated dosimeters was the end of
the irradiation. The dose to each dosimeter is displayed on the figure. A dotted line is given as a visual aid to the time
dependenc-e of the four un-irradiated dosimeters for which time zero was the time they were cut from the sheet. A jump in the
rate of change, most easily seen at 65 days for the dosimeter that received 14.52 Gy, occurred over the same dates for all the
dosimeters, time zero having ctifferent dates for the different dosimeters. (From Reference [11]. With permission.)
The change in OD was measured vs. time for a number GafChromic MD-55 is composed of seven layers as
of dosimeters, some un-irradiated and others irradiated illustrated in Figure 5.27. Either outer Mylar layer could
to a variety of doses up to 14.52 Gy. Over this dose range, be pulled off the gel by flexing the dosimeter under liquid
I Gy corresponds to about 0.1 I OD units. Assuming 0.1 I nitrogen and pulling the Mylar layer off or, alternatively,
OD unit,s equals lGy, the rate of change of OD is shown in by soaking the dosimeter in water overnight and pulling
Figure 5.26 as the equivalent rate of change in cGy day*l. it apart at the Mylar/gel interface.
(DOD is the change in ab,sorbance due to polymer In order to optimize the measurement sensitivity and
formation.) thus improve precision, Reinstein et a1. [12] described a
Film Dosimetry 3J 9

the light source it(A) with the absorption spectra of

GafChromic film, the intensity reaching the detector can
be written as

(5.15)

where AG p(h, D) is the absorbance of GafChromic film

at wavelength ñ, irradiated to a dose D. Optical density
as a function of dose [OD(D)] for GC film was calculated
by combining the equations above to yield

D
OD(D) — log , 2o( )10 ’“”"’10* " ""’ ' x
(5.16)

FIGURE- 5.27 A schematic view of the structure of

GafChromic MD-55, indicating the seven layers and their By subtracting the optical density of an un-irradiated
thicknesses. The uncertainty in the measurement of the sample of GafChromic film, the net optical density (NOD)
thicknesses was estimated t‹a be about *0.001 mm Y‹ r the as a function of dose was obtained (dose response curve)
hard Mylar layers and *0.002 mm for the soft, gel, and glue for each densitometer light source. Note that the model
layers. (From Reference [11]. With permission.) assumes a uniform response by the detector for the range
of wavelengths measured. The transmission spectra for
method to calculate the dose response curves (net optical the two band-pass filters used in our measurements with
density at a given wavelength or spectrum vs. absorbed the filtered LED densitometers are shown in Figure 5.28.
dose) for different densitometer light ,sources u,sing Absorption spectra of GC film exposed to total doses 0,
measured GC film absorption spectra. Comparison with 10, 20, 30, 50, and 100 Gy and measured at 31.9 days
measurements on the latest ver.sion of GC fi1m (Model post-irradiation are shown in Figure 5.29. The minor and
MD-55-2) u.sing four types of densitometers [He-Ne laser, major absorption peaks can clearly be seen. At 30 Gy
broadband (white light) densitometer, and two LED (red- exposure, these peaLs are centered at 614 nm and 674 nm,
light) filtered den- sitometers] confirm the accuracy o1 this respectively, with peaks shifting to lower wavelengths
predictive model. They found that an LED (red-light) with increasing dose.
,source with a narrow band-pass filter centered at 671 nm Figure 5.30a plots the predicted dose response curves
near the major absorp- tion peak achieves nearly the at fixed wavelength,s 400 rim, 510 nm, 632.5 nm, 650 rim,
maximum possible sensitiv- ity (almost four times more and 671 nm as calculated from the measured absorption
sensitive than He-Ne laser, 632.S nm) and may be suitable spectra of GC film (MD-55-2). It is observed that the dose
for in vivo dosimetry. response curve measured at 671 nm (near the major peak)
For LED light source coupled to a band-pass filter, is the most sensitive with a DNODI = 14 Gy. The least
the resulting intensity as a function of wavelength after sensitive dose response curve is at 400 nm (not near either
having passed through the filter is of the absorption peaks), where a net optical density of
1 is not achieved, even for a dose as high as 100 Gy. Of
the fixed wavelengths, the dose response curve at 650 nm
(5.14) (on the left shoulder of the major peak) exhibits the great-
est linearity, with a DNODI = 34 Gy. Also shown i.s the
where f (A) is the intensity of the light source (LED) at dose response curve for a fixed wavelength of 632.5 rim,
wavelength A incident on the filter; I (h) i,s the intensity which can be compared with the measured data from the
after having passed through the filter; and A¿„„(h) is the He-Ne laser densitometer. The predicted and measured
absorbance of the filter at wavelength A. For the case of results show excellent agreement (DNODI = 56 Gy for
the broadband (white light) light source, I0(h) was deter- both). [12]
mined from the temperature of the tungsten filament, an The dose response of high-sensitivity GafChromic
assumed blackbody radiation spectrum, and, for mathe- film to photons from'2^ I seeds for doses up to 200 Gy
matical convenience, A¿ „{h) — 0 (no filter). By coupling
320 Radiation Dosimetry: instrumentation and Methods

FIG URE 5.20 Transmission spectra for the 660 rim filter (dashed line) and 621 nm filter (solid line) measured using the Du 640B
spectr‹ameter. (Frram Reference [12]. With permission.)

FIG URE 5.29 Absorbed spectra of GafChromic film (MD-55-2) for the dose levels 0, 10, 20, 30, 50, arid 100 Gy. The spectra were
measured on a Bec-kman DU 640B spectrometer. (From Reference [12]. With permission.)

was established by Chiu-Tsao et al. [13] The net optical increa.sing dose. The values with the LKB were about
density was found to be a power function of dose with twice those with the Macbeth, depending on the dose.
exponents of 0.858 and 0.997 1or the Macbeth and LKB This means that the film sensitivity measured with red
densitometers, respectively. Film sensitivity with the LKB light from the He-Ne laser is higher than that obtained
densitometer was about double that with the Macbeth with the broadbeam spectrum. The data obtained with the
densitometer. Macbeth densitometer were fitted to a power function of
The dose-response curves of the high-sensitivity dose, with exponents of about 0.858, 0.839, and 0.849
GafChromic films were plotted for '2^I, '*7Cs, and ^0Co for '2'I, "7Cs, and ^0Co, respectively. As for the LKB
and shown in Figures 5.31a, b, and c, respectively, for densitometer, the curves were linear for ’" 7Cs and ^0Co.
doses up to 200 Gy. The curves for the same radionuclide
The curve for '2'I, however, was better-fitted to a power
obtained with the LKB and Macbeth densitometers were
function of dose with a power of 0.997, i.e., essentially
compared to the figure. The net OD increases with
linear.
Film Dosimetry 321

FIG URE 5.30 (a) Predicted dose response curves when mea-
sured at fixed wavelengths 400 rim, 510 rim, 632.8 nm, 650
nm, and 671 rim for GafChromic film (MD-55-2), determined
from its absorption spectra using a Beckman DU 640B
spectropho- tometer at 31.9 days post-irradiation (solid
curves). Measured data (A) using the He-Nc laser is also
shown for comparison with the predicted dose response curve
at fixed wavelength of
632.8 rim. Three samples were given the same total dose of 30
Gy and used to check for measurement reproducibility. (b)
Pre- dicted dose response curves (solid curves) for a tenable
light source to the minor peak (A), major peak (B), and for
valley between peaks (C), and for complex light s‹surces: an
LED light source coupled to a 660-rim band-pass filter (D), an
LED light source coupled to a 671-rim filter (E), and a
broadband light source (F). Measured dose response curves
are also shown for comparison, namely a 660-rim filtered FIGURE 5.34 The dose-response curves of the high-
densitometer (A), 671-nm filtered densitometer (O) and a sensitivity GafChromic film for three radionuc1ides,'2^I, 13'Cs,
broadband densitometer (x). Three samples were given the and ^"Co, in a, b, and e, respectively. The square and circle
same total dose of 30 Gy and were used to check for symbols are for the reading using Macbeth and LKB
measurement reproducibility. (From Reference [12]. With densitometers, respectively. The solid o-urves represent the c-
permission.) orresponding function fit. (From Reference [13]. With
permission.)
322 Radiation Dosimetry: instrumentation and Methods

FIGURE- 5.32 Dose response curves for radiochromic

FIGURE 5.34 Family of dose-response curves for silver
film. (From Reference [14]. With permission.)
halide film. Each curve is for one x-ray beam (tone equivalent
pht›ton energy). (From Reference [14]. With permission.)

FIGURE- 5.33 Family of dose response curves for radiochromic

film. (From Reference [14]. With permission.)
FIGURE 5.35 Family of dose-response curves for LiF
TLD chips. Each curve is for one type of x-ray beam (one
The dose response of the radiochromic film for 4-MV equivalent photon energy). (From Reference [14]. With
x-rays is shown in Figure 5.32. [14] The x-rays were permission.)
produced by a Clinac 4 linear accelerator, and the film
wa,s irradiated at a depth of 1.25 cm in an acrylic phan-
doses are about 250 times the doses used for the silver
tom set up at an SSD of 41 cm. The dose rate was about
halide film curves (Figure 5.34) and about 600 times the
900 cGy/min. The lower curve shows the results from
doses used for the LiF chip curves (Figure 5.35). The
optical density measurements using broadband absorp-
curves show linearity of dose response, and the similarity
tion, and the upper curve shows the results of measure-
of slopes shows that the dependence on photon energy is
ments using 632.8-rim light.
slight. The optical densities are for 632.8-rim light.
A family of dose-response curves for radiochromic
An “improved film” was developed with higher sensi-
film for different photon energie,s i,s ,shown in Figure 5.33.
tivity and better uniformity than the model MD-55 films
These curves are in the low-dose range for this film,
by Meigooni et al. [I 5] The improved film is a
although the
multilayered structure composed of a nominal 30-pm
thickness of the
Film Dosimetry 323

FIGURE- 5.36 Sensitometric curves of the improved radiochro-

mie film (solid symbol) compared with that of MD-SP film
(open symbol). These results were obtained by irradiating the
films to doses ranging from 1—72 Gy of gamma rays from a
^"Co tele- therapy unit. (From Reference [15]. With
permission.)

sensitive material, sandwiched between two pieces of 75-pm- FIGURE 5.37 Relative optical density of a uniformly irradiated
thick polyester base material and two pieces of 13 jzm improved radiochromic film (a) measured in one direction along
the film and (b) measured along the orthogonal direction. (From
lam- inated material, yielding a total thickness of 206 pm.
Reference [15]. With permission.)
In contrast, MD-55 film is composed of a layer (23-pm-
thick) of sensitive material on one side of a 100-micron
improved radio-chromic film. A similar result was
polyester base. These films are colorless and transparent
obtained at a distance of 5 cm from the beam axis, along
before heing exposed to ionizing radiation. A,s they are
the same direction of the film. However, along the per-
irradiated, their color changes to blue, due to the
pendicular direction (Figure 5.37b), a non-uniformity of
polymerization of radio- chromic dye. The darkne.ss o1 the
up to 159c wa,s observed. This non-uniformity is not con-
blue color depends on the absorbed dose and can be
sistent from one sample of film to another.
measured with a laser densitom- eter with a wavelength
Figure 5.38 shows the variation of relative optical
ranging from 610—675 nm.
density of the improved radiochromic film and MD-55
Figure 5.36 shows the measured sensitometric curves
film as a function of time after irradiation for absorbed
of the improved radiochromic film and MD-55 film as a
dose as shown in the figures.
function of absorbed dose. This figure indicates that the
Evaluation of the influence of irradiation temperature
improved film requires about 409c less dose than MD-55
differences on the response of three widely used
film to yield an optical density of 0.5. This difference is
radiochro- mic filin dosimeters was made by McLaughlin
reduced to about 30'fa for an optical density of about 1.5.
et al. [16] Freestanding nylon base film (mean thickness
The combination of all sources of uncertainties in these
48 pm) FWT-60-00, thin-coated (-7 pm) sensor on 100-
measurements is estimated to be less than *5' a. This
pm poly- ester base GafChromic DM-1260 dosimetry
uncertainty includes the reproducibility of the la,ser scan-
media, and freestanding polyvinylbutyral base film (mean
ning system, uncertainty of determination of absorbed
thickness 23 pm) Risk B3 were investigated.
dose given to the film, and standard deviation of the
The irradiations were made with gamma-rays from two
optical density measured along a path of about 5 mm in a
^ Co sources, one at an absorbed dose rate of 12.2 kGy/h
uni- formly exposed film.
and the other at 30.0 kGy/h. During irradiation the films
The uniformity of the improved radiochromic film and
(three films for each irradiation) were held between two
MD-55 film along two orthogonal directions was deter-
electron equilibrium layers of 3-mm-thick polystyrene and
mined by measuring the optical density o1 a film
sealed at 509c relative humidity in a triple layer pouch of
irradiated with a uniform radiation field. Figure 5.37a
shows very good uniformity (within 49c) along one
direction of the
324 Radiation Dosimetry: instrumentation and Methods

FIGURE 5.39 Temperature dependence, relative to that at

20°C, of the gamma-ray response of two batches of radiochromic
film FWT-60-00. Batch 8W9 was irradiated at three
absorbed doses (6, 15, and 30 kGy) and batch 1D0 was
irradiated at 15 kGy. The absorbances per unit thickness (AA
mm*') were mea- sured at 605-nm wavelength. (From
Reference [16]. With per- mission.)

emitting laser diode (675.2 nm). The effective sensitivity

of the film, compared with a helium-neon laser densito-
meter (632.5 nm), is increased approximately threefold.
Figure 5.42 shows the optical density of the radiochro-
mic film depending on the wavelength of the analyzing
FIG URE 5.38 Variation of the optical density with time
beam after irradiation with four different doses. [18] In
after irradiation for the improved radiochromic film (solid
the helium-neon laser densitometer, the film is scanned
symbol) compared to MD-55 film (open symbol). These
results were obtained by Simultaneous irradiation of pieces of with an analyzing beam and the transmission is measured
each film-type absorbed doses of 5, 16.92, and 33.85 Gy using by a photomultiplier. The emission wavelength (632.8 nm)
a °'Co telether- apy unit. (From Reference [15]. With of the analyzing beam lies in the valley of the film's
permission.) ab.sorption spectrum. Since the absorption spectrum of the
radiochromic film contains two peaks centered at approx-
imately 615 nm and 675 nm, it has been suggested [17]
30-pm-thick polyethylene—aluminum foil laminate. Each that significant improvement in response might be
packet was conditioned 1or 30 min at each irradiation achieved by tailoring the wavelength of the incident den-
temperature before irradiation, then irradiated, and then sitometer light to one of these peaks. Densitometers with
stored at room temperature immediately after irradiation. a light-emitting diode (LED) (660 nm, spectral line width
Figures 5.39, 5.40, and 5.41 show, respectively, the >30 nm) and with an LED coupled to a band-pass filter
temperature dependence of the gamma-ray respon,ses of (670 nm, bandwidth 11 nm, GatChromic densitometer)
the three mo,st recent batches of radiochromic films, have been presented, but their spectral line width is too
FWT- 60-00 (batches 8W9-1989 and 1D0-1990), broad to match the absorption maximum exactly, and the
GafChromic 1260 (batch 09031501-1989), and Ri.st B3 improvement in response is reduced.
(batch 85-128- 1959). In all three ca,ses, it i,s ,seen that a The radiochromic film's sensitive layer consists of
different trend occurs at different absorbed doses, microcrystals of radiation-sensitive monomer uniformly
particularly at the higher dispersed in a gelatin binder. When the microcrystals
temperatures. are exposed to ionizing radiation, polymerization occurs
A technique of readout was developed by Kellermann and the polymers alter the crystal color to various
et al. [17] to measure the optical density distributions of shades of blue. The optical density increases
the film in purely directed light. This technique imple- continuously with the absorbed dose. The optical
ments radiochromic film dosimetry near the film's density depends neither
absorption maximum by u,sing a single-mode top-surface
Film Dosimetry 325

FIGURE- 5.40 Temperature dependence, relative to that at 20°C, of the gamma-ray response of radiochromic film
GafChromic Dcasimetry Media DM-1260 batch 09031501 irradiated at two absc›rbed doses. The wavelengths of absorbance
measurements:
•—650 rim; + —600 rim; X 510 rim; 0 —400 rim. (From Reference [16]. With permission.)

FIGURE- 5.44 Temperature dependence of the gamma-ray response o1 radiochromic film Risk B3, batch 85-128, irradiated at
two absorbed doses. The two right ordinates show the response relative to that at 20°C. The absorbanees per unit thickness
(AA mm 1) were measured at 554-nm wavelength. (From Reference [16]. With permission.)

on photon energy in the range 0.1 to 20 MeV nor on orthogonal to it. This would result in a dose error of
the energy of secondary electrons in the range 0.01 to 10' a or 2.5' a.
20 MeV. A laser diode with an emission wavelength of 675.2
nm GafChromic MD-55T^ irradiated in a uniform radia- was chosen by Kellermann et a1. as a light source to trac-
tion field showed a non-uniformity in optical density of silluminate the radiochromic film. At 675.2 rim, near the
up to 159c in one direction and within 49c in the direction film's absorption maximum, a dose gradient in the
radiation
326 Radiation Dosimetry: instrumentation and Methods

FIG URE 5.42 Optical density of the radiochromic film (Gafchromic MD-55 , Model No. 37-041), depending on the
wavelength of the analyzing beam after irradiation with four different doses. [18] The wavelength of the helium-neon, laser
(632.8 nm) lies in the valley of the film’S absorption spectrum, while the wavelength of the laser diode (675.2 nm) used by
Kellermann et a1. is near the film's absorption maximum. (From Reference [17]. With permission.)

FIGURE- 5.43 The schematic measurement setup is sketched. The laser light transmits the film and is sent via art optical fiber to
a solar cell, where it is transformed to an electrical signal. This is measured with a lock-in amplifier. The light of the edge
emission is used as a reference Signal. (From Reference [17]. With permission.)
field leads to a change in optical density that is up to three
After irradiation, the optical density of radiochromic
times larger than the change that i.s caused by the same
film increases logarithmically. Therefore, the optical den-
dose gradient in the valley (632.8 nm) of the film's absorp-
sity measurements with the densitometry system are done
tion spectrum. The laser light propagating in the active
four days after the end of irradiation, by which time the
region excites a transverse electrically polarized (F6 0) sur-
changes in optical density are small.
face mode in the waveguide structure on the top of the
The calibration curv hange in optical density at
laser diode through grating coupling. The surface mode
675.2 rim vs. radiation dose-is shown in Figure 5.44. The
couples both into the vacuum light cone, resulting in top-
curve rises rapidly, but the upward gradient becomes
surface emi.ssion with narrow beam divergence, and back
smaller with higher doses. In comparison, the optical
to the active region, leading to a gain mechanism and thus
density vs. radiation dose at 632.8 rim is flat but
to single-mode emission.
approximately linear.
The film is fastened (with double-.sided adhesive tape)
[19] At 675.2 rim, a dose gradient from 0 Gy to 20 Gy in
to a stage that allows precise movements along the scan-
the radiation field leads to a change in optical density that is
ning direction (and orthogonal to it) in 50-pm steps along
3.I times larger than at a wavelength of 632.8 rim. This
a distance of 75 mm. The schematic of Figure 5.43 dis-
ratio decreases to 1.9 with a dose gradient from 20 Gy to
plays the measurement setup.
40 Gy. Film digitizers are, compared with mechanical
scanners, much faster, while their spatial resolution is
even higher.
Film Dosimetry 327

FIGURE- 5.44 Calibration curve—change in optical density at 675.2 nm vs. radiation. The curve rises rapidly, but the upward
gradient becomes smaller with higher doses. In comparison, optical density vs. radiation dose at 632.8 rim is flat but
approximately linear. (From Reference [17]. With permission.)

Most commercially available digitizers are based on either

covering the 1.5 to 50-Gy dose range, when used with the
a scanning He-Ne laser beam or a one-dimensional charge-
radiochromic densitometer. Higher doses (100 Gy and
coupled device (CCD) combined with a standard fluores-
greater) can be measured using other commercially avail-
cent lamp. Mer,sseman and De Wagter [20] investigated a
able densitometers. MD-55 has a higher sensitivity than
CCD-based digitizer.
HD-810 and was developed for use in external-beam
The principles of the digitizer are as follows. A broad radio- therapy and brachytherapy applications. MD-55 is a
light beam from the fluore.scent lamp (Philips F17T8/ lam- inated film composed of two pieces of 2.65-mil
TL84l ) crosses the light-diffu,sion plate and travels polyester base, each with a nominal 15-micron-thick
through the scrolling film to be digitized. Two fixed ver- coating.
tical and two movable horizontal shutters subsequently HD-810 (Model 37-040) films are low-sensitivity
collimate this light beam to a narrow rectangular field that 8“ X 10“ clear film sheets suitable for high-dose beam
covers the film lengthwise. The transmitted light is sub- pro- filing and dose-mapping applications in the 10 to 300-
sequently reflected by a stationary mirror, focused by the Gy dose range, when used with the radiochromic
40-mm lens and finally detected by the linear CCD ele- densitome- ter. Higher doses (up to 2500 Gy) can be
ment. The CCD detector reads line per line at 300 dot,s measured using other commercially available
per inch (dpi) over its full length of 35.6 cm (14 in). In densitometers. This media has a nominal 7-micron-thick
order to achieve a 150, 75, and 60 dpi resolution, the film radiation-sensitive layer on a 3.9 mil polyester base.
digitizer combine.s data from nearby pixels. A 12-bit GafChromic Dosimetry Media is composed of
analog-to-digital converter digitizes the ,signal. Conver- materials with low atomic numbers, so the fi1m can be
sion tables (built-in or loaded from the host computer) are considered ti,ssue-equivalent for most radiation therapy
allowed to further translate the data. The resulting data applications. Additionally, tl s material has low sensitivity
matrix is transferred to the host computer (Pentium PC) to ambient room light, which simplifies handling
using an SCSI interface. procedures and enhances image stability. Its response is
Nuclear Associates' GafChromic Dosimetry Media dose rate indepen- dent; it is linear dose response and
is a radiation-sensitive imaging film designed for use in energy independence above 100 keV. [21]
medical and industrial quality control measurement
applications. The media is colorless and grainless and
offers a very high spatial resolution (1200 LP/mm), mak- TECHNICAL DATA
ing it invaluable for measuring dose distribution around MD-55: Polyester Base: 67-pm; Sensitive Layer: 15-pm;
small brachytherapy source,s and ,stereotactic radiosur- Pre,ssure-Sensitive Adhesive Layer: 44.5 pm; Polyester
gery fields. [21] GafChromic Dosimetry Media MD-55 Ba,se: 25 pm; Pressure-Sensitive Adhesive Layer: 44.5
(Model 37-041) comprises 5“ X 5“ clear film sheets that pm; Sensitive Layer: 15 pm; Polyester Base: 67 pm.
are suitable for medium-dose measurement applications HD-810: Sensitive Layer: 7 jzm; Adhesive Layer: 1.5 join;
Conductive Layer: 0.05 pm; Polyester base: 99 jzm.
328 Radiation Dosimetry: instrumentation and Methods

FIG URE 5.45 GafChromic sensitivity to energy composed to that of silver based film on LiF TLD. (From Reference [21].
With permission.)

FIG URE 5.46 Response of radioc-hromie film of tetrazolium chloride in polyvinyl pyrrolidone (in terms of increase in
absorbance measured at 522 rim wavelength) as a Junction of the gamma ray absorbed dose. The inset shows the increase in
the absorption spectral amplitudes with dose. (From Reference [22]. With permission.)

37-041 MD-55 GAFCHROMIC Dosimetry Media; doses consists of a film or gel cast from an alcohol solu-
Package of Five 5” X 5” Sheets, Weight: .10 lb (.04 kg). tion of blue tetrazolium or triphenyl-tetrazolium chloride
37-040 HD-810 GAFCHROMIC Dosimetry Media:
combined with a polymer matrix material (e.g. polyvinyl
Package of Five 8” X 10” Sheet,s, Weight: .50 lb (.24 kg)
alcohol, polyvinyl pyrrolidone), McLaughlin [22]. Figure
The GAFCHROMIC sensitivity to energy compared
5.46 shows the radiation response properties of one exam-
to that of silver based film is shown in Figure 5.45.
ple of this film system, in this case with a thickness of
Another promising thin film polymer dosimeter 1or high
30 pm.
Film Dosimetry 329

REFERENCES
1.
2. Cheng, C-W. and Das, I. J., Med. Phy,s., 23, 1225, 1996. 13. Chin-Tsao, S-T. et al., Med. Phy,s., 21, 6fi1, 1994.
Burch, S. E. et al., filed. Phys., 24, 775, 1997. 14. Mueneh, P. J. et al., Med. Phys., 18, 769, 1991.
Robar, 1. L. and Clark, B. G., Med. Phys., 26, 2144, 15. Meigooni, M. F. et al., Med. Plus., 23, 1583, 1996.
4. 1999. 16. McLaughlin, W. L. et al., in Pmc. High Dose
El-Khatib, E. et al., Med. Phys., 19, 317, 1992. Dosim- etry far Radiation Pmtectian, IAEA, 1991,
6. Stern, R. L. et al., Med. Phys., 19, 165, 1992. 305.
7. Novotny, 5• et a1., Phy,s. Med. Bial., 42, 1277, 1997. 17. Kellermann, P. O. et al., Phys. Med. Biol., 43, 2251,
Stevens, M. A. et a1., Phy,s. Idled. Biol., 41, 2357, 1998.
1996. McLaughlin, W. L. et al., Nucl. Iii,st. Beth., 18. Ertl, A. et al., Phys. Med. Bial., 43, 1ñ67, 1998.
9. A302, 165, 1991. 19. Meigooni, A. C. et al., Med. Phy,s., 23, 1503, 1996.
Vatnitsky, S. M. et al., Phys. Med. Biol., 42, 1887, 20. Mersseman, B. and De Wagter, C., Phys. Med. Bial.,
10. 1997. 43, 1803, 1998.
11. Zhu, Y. et al., Med. Phys., 24, 223, 1997. 21. Nuclear Associates, Diagnastic Imaging and Radiatian
\2. Klassen, N. V. et al., filed. Phy,s., 24, 1924, Therapy Catalog, 1999.
1997. 22. McLaughlin, W. L., in Proc. High Do,se Dosimetry
Reinstein, L. E. et aI., Med. Pliy,s., 24, 19.35, 1997. [or Radiation Pmtectiori, IAEA, 1991, 3.